Fake food photos are everywhere: impossibly perfect recipes on social media, restaurant menus illustrated by AI, ads for dishes that don't exist. At a glance, an AI-generated burger or dessert often passes for real. But a few habits are enough to catch the fake.

Why food fools the eye

Food is easy ground for AI: appetising colours, soft light, blurred background — all things generators handle well. But food also has a physical logic — texture, cooking, gravity — that AI still reproduces poorly.

5 tells that give away an AI dish

1. Impossible textures

Cheese dripping into thin air, rice grains fused together, a bread crust with no coherent crumb. Zoom in: the matter doesn't "hold" like in reality.

2. Inconsistent details

A three-pronged fork, a spoon passing through the plate, two pieces of the same food melting into each other. AI polishes the centre of the dish and neglects the edges.

3. Gibberish text

A label, a menu, packaging: the letters turn to mush. Whenever there's text in the image, that's a classic AI weak point.

4. The "plastic" look

An overly uniform gloss, perfect reflections everywhere, food that looks like a 3D model rather than a real dish. "Too smooth" should raise a flag.

5. Illogical portions and reflections

A glass whose reflection matches nothing, a shadow that ignores the light, quantities that defy physics. Look at the whole scene, not just the food.

Careful: real professional food photos are gorgeous too. "It's too beautiful" is no longer enough — rely on texture and logic inconsistencies.

Train your eye: Food mode

Detection is learned by practice. In our Food mode, a dish appears each round: decide whether it's a real photo or an AI image. The real ones come from food photographers; the fakes are generated for the occasion.

Real photo or AI dish? How many can you spot in a row?

Play Food mode →

In short

AI makes food look appetising, but still trips on matter and logic. Hunt for impossible textures, inconsistent details, gibberish text and the plastic look. And practise: after a few games, the eye spots the anomaly before you even know why.